Origin and evolution of Pliocene–Pleistocene granites from the Larderello geothermal field (Tuscan Magmatic Province, Italy)

Abstract

Extensive, mainly acidic peraluminous magmatism affected the Tuscan Archipelago and the Tuscan mainland since late Miocene, building up the Tuscan Magmatic Province (TMP) as the Northern Apennine fold belt was progressively thinned, heated and intruded by mafic magmas. Between 3.8 and 1.3 Ma an intrusive complex was built on Larderello area (Tuscan mainland) by emplacement of multiple intrusions of isotopically and geochemically distinct granite magmas. Geochemical and isotopic investigations were carried out on granites cored during drilling exploration activity on the Larderello geothermal field. With respect to the other TMP granites the Larderello intrusives can be classified as two-mica granites due to the ubiquitous presence of small to moderate amounts of F-rich magmatic muscovite. They closely resemble the almost pure crustal TMP acidic rocks and do not show any of the typical petrographic features commonly observed in the TMP hybrid granites (enclaves, patchy zoning of plagioclase, amphibole clots). On the basis of major and trace elements, as well as REE patterns, two groups of granites were proposed: LAR-1 granites (3.8–2.3 Ma) originated by biotite-muscovite breakdown, and LAR-2 granites (2.3–1.3 Ma) generated by muscovite breakdown. At least three main crustal sources (at 14–23 km depth), characterized by distinct ε Nd( t) and 87Sr/ 86Sr values, were involved at different times, and the magmas produced were randomly emplaced at shallow levels (3–6 km depth) throughout the entire field. The partial melting of a biotite-muscovite-rich source with low ε Nd( t) value (about −10.5) produced the oldest intrusions (about 3.8–2.5 Ma). Afterwards (2.5–2.3 Ma), new magmas were generated by another biotite-rich source having a distinctly higher ε Nd( t) value (−7.9). Finally, a muscovite-rich source with high ε Nd( t) (about −8.9) gave origin to the younger group of granites (2.3–1.0 Ma). The significant Sr isotope disequilibrium recorded by granites belonging to the same intrusion is interpreted, as due to the short residence time of magmas in the source region followed by their rapid transfer to the emplacement level. Partial melting was probably triggered by multiple, small-sized mafic intrusions, distributed over the last 3.8 Ma that allowed temporary overstepping of biotite- and muscovite-dehydration melting reactions into an already pre-heated crust. Dilution in time of the magmatic activity probably prevented melt mingling and homogenization at depth, as well as the formation of a single, homogeneous, hybrid pluton at the emplacement level. Moreover the high concentrations of fluxing elements (B, F, Li) estimated for the LAR granites modified melt properties by reducing solidus temperatures, decreasing viscosity and increasing H 2O solubility in granite melts. The consequences were a more efficient, fast, magma extraction and transfer from the source, and a prolonged time of crystallization at the emplacement level. These key factors explain the long-lived hydrothermal activity recorded in this area by both fossil (Plio-Quaternary ore deposits) and active (Larderello geothermal field) systems.